From Dolly the sheep to embryo selection: reflecting on a decade of progress in genetics and assisted reproduction

Some things never change. When the first issue of BioNews appeared on 29 March, 1999, the second item announced that an Italian doctor was claiming to have helped ten infertile men to become fathers, by cultivating their sperm in the testicular tissue of rats. Just this morning, I was asked by The Times's foreign desk for my views on another claim by the self-same doctor - this time, that he has cloned three children, all nine years old and living healthily in Eastern Europe.

The past decade has clearly done little to dull Severino Antinori's appetite for self-publicity without evidence. But so much else has moved on.

In 1999, the first human embryonic stem cells (ES cells) had been derived just a year earlier, and Dolly celebrated her third birthday. This year has already brought the first patient study of an ES cell therapy, and improvements in the production of iPS (induced pluripotent stem) cells that might eventually circumvent the need for embryos.

The first issue of BioNews declared the Human Genome Project 'ahead of schedule', but it was still years away from completion at an eventual price tag of around $4bn. Last month, Complete Genomics announced it will offer human genome sequencing for $5,000 from June, and companies like 23andMe are selling genotyping data direct to consumers.

Reproductive medicine has also advanced, if not in quite such eye-catching fashion. Game-changers such as ICSI (intracytoplasmic sperm injection) and PGD were already with us in 1999 - but they are now well-integrated into clinical practice. IVF, and demand for IVF, is now seen as routine. And second-generation PGS techniques such as CGH and Parental Support are promising to take embryo selection to a new level.

It is always difficult to predict the future of science: if we knew which hypotheses were going to be right, and which drugs and therapies would work, there would be no need to do any experiments. But I expect the progress BioNews reports in its second decade to exceed even the advances that marked its first ten years.

What is more, that progress will pose direct challenges for medicine and society - perhaps more quickly than many people imagine. Scientists, clinicians and ethicists are already starting to consider how we should respond. But there are few signs of similar reflection among the politicians who will have to manage these developments, or the wider public who will be affected.

This applies to stem cells and to assisted reproduction, but especially to genomics. It is hard to overestimate the pace at which this is moving forward. The price of genome sequencing is moving inexorably downwards. It is already possible to genotype 500,000 SNPs for £300 a head - that is what 23andMe does - and a complete genome for a similar price is not going to be far away. The main barriers to providing this service for everyone are not going to be financial, but medical and social.

Is there actually any clinical value in knowing which common genetic variants a patient carries? Assuming there is - and very few common SNPs have yet been clinically validated as predictive markers - how will people respond to learning they have a raised or lowered background risk of heart disease? Will they become fatalistic, complacent, or scared to the point of anxiety? Will people worry about their genetic privacy, and avoid tests that might be helpful for fear of discrimination by insurers or employers? Can such privacy even be protected, when we leave DNA behind us everywhere we go? These are not just questions for scientists and ethicists - they are questions for us all.

Then there are the public policy implications of interpreting genomic information. So far, only a very few people have been genotyped by the direct-to-consumer companies, yet clinical geneticists, GPs and counsellors are already reporting visits from patients who want to know what their results mean. Fast forward to when most of us have had our whole genomes sequenced. Who is going to interpret the results? GPs will do their best, but few are trained to tease out risk profiles from SNP data.

Pharmacogenomics poses a further challenge. As we understand more about how genetic variation affects cancer and drug metabolism, we can expect the development of many more targeted drugs, matched to patients with particular genotypes or genetic sub-types of tumour. The first of these drugs, such as Herceptin, have proved however to be extremely expensive - how will the NHS be able to pay? The costs should come down as pharmaceutical companies use genetic data to optimise patient populations for clinical trials, but that will require a rethink of the appropriate regulatory regime

In his speech on science at the end of last month, Gordon Brown predicted a great future for genomic medicine. 'It is hard not to see how -- with the appropriate safeguards in place -- the potential of genomic information might prove to be a 21st century moment, a vast expansion of the boundaries of scientific understanding that holds breath-taking possibilities for the future effectiveness of medicine in Britain and across the world,' he said.

I think the Prime Minister is probably right. But there is little evidence yet that his government - or any of the other parties, for that matter - have yet started to think deeply about the many new questions that this expansion will pose.

This suggests a job beyond reporting for BioNews over the next decade. Its editors, and especially, its readers, need to drive these issues high onto the political agenda.